Registration of Time Series
If the sample is characterized by movement during the acquisition the time series can be corrected by aligning successive frames to the first.
The method assumes a 2D time series as input
- Sends the time series to imageJ where the plugin StackReg is applied with a 'rigid' transformation model.
- Rigid: only translation or rotations i.e. no scaling
- Attention: make sure to allocate enough Java Heap Space in Matlab to be able to process the image
- A Pyramid Approach to Subpixel Registration Based on Intensity P. Thévenaz, U.E. Ruttimann, M. Unser IEEE Transactions on Image Processing, vol. 7, no. 1, pp. 27-41, January 1998.
- StackReg website
Parameter: EpiTools's registration [beta]
- This experimental registration method aligns the images in two steps:
- registration based on big features
- correction based on small features